| *** phalmos has quit IRC | 00:03 | |
| *** lbragstad has joined #openstack-ansible | 00:06 | |
| *** thorst has joined #openstack-ansible | 00:16 | |
| *** yifei has joined #openstack-ansible | 00:21 | |
| *** thorst has quit IRC | 00:21 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Disable offloading in test by default https://review.openstack.org/523541 | 00:25 |
|---|---|---|
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Disable offloading in test by default https://review.openstack.org/523541 | 00:30 |
| *** vnogin has joined #openstack-ansible | 00:44 | |
| *** chyka has quit IRC | 00:47 | |
| *** thorst has joined #openstack-ansible | 00:49 | |
| *** thorst has quit IRC | 00:54 | |
| *** vishwanathj has joined #openstack-ansible | 00:57 | |
| *** vnogin has quit IRC | 00:57 | |
| *** thorst has joined #openstack-ansible | 00:58 | |
| *** thorst has quit IRC | 00:59 | |
| *** thorst has joined #openstack-ansible | 01:06 | |
| *** hw_wutianwei has joined #openstack-ansible | 01:09 | |
| *** thorst has quit IRC | 01:21 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Fix container restart handler https://review.openstack.org/523642 | 01:33 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Disable offloading in test by default https://review.openstack.org/523541 | 01:38 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Disable offloading in test by default https://review.openstack.org/523549 | 01:39 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Gather more logs when running tests https://review.openstack.org/523647 | 01:42 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts master: Add more testing and increase build timeout for suse/cent https://review.openstack.org/523525 | 01:47 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server master: SUSE: Apply workaround for mariadb-10.2 https://review.openstack.org/523430 | 01:50 |
| *** thorst has joined #openstack-ansible | 01:51 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server master: Implement a proper WSREP check for galera https://review.openstack.org/520665 | 01:52 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Fix container restart handler https://review.openstack.org/523642 | 01:56 |
| *** thorst has quit IRC | 01:57 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server master: Tune-up the galera role for efficiency https://review.openstack.org/466827 | 01:59 |
| *** thorst has joined #openstack-ansible | 02:00 | |
| *** hamzy has joined #openstack-ansible | 02:03 | |
| *** thorst has quit IRC | 02:05 | |
| *** chyka has joined #openstack-ansible | 02:17 | |
| *** tnogisto2 has joined #openstack-ansible | 02:19 | |
| *** chyka has quit IRC | 02:21 | |
| *** woodard has joined #openstack-ansible | 02:21 | |
| *** tnogisto has quit IRC | 02:22 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Fix container restart handler https://review.openstack.org/523642 | 02:31 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Gather more logs when running tests https://review.openstack.org/523647 | 02:33 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Disable offloading in test by default https://review.openstack.org/523541 | 02:34 |
| *** woodard_ has joined #openstack-ansible | 02:41 | |
| *** woodard has quit IRC | 02:41 | |
| *** dave-mccowan has quit IRC | 02:42 | |
| *** threestrands has joined #openstack-ansible | 03:05 | |
| *** threestrands has quit IRC | 03:05 | |
| *** threestrands has joined #openstack-ansible | 03:05 | |
| *** thorst has joined #openstack-ansible | 03:05 | |
| *** ashak has quit IRC | 03:05 | |
| *** thorst has quit IRC | 03:10 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Fix container restart handler https://review.openstack.org/523642 | 03:13 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_barbican master: Update paste, policy and rootwrap configurations 2017-11-28 https://review.openstack.org/523668 | 03:19 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Update all SHAs https://review.openstack.org/522850 | 03:23 |
| cloudnull | anyone around the might be able to give this a push through https://review.openstack.org/#/c/523430/ | 03:25 |
| cloudnull | fixes opensuse builds for galera | 03:25 |
| *** thorst has joined #openstack-ansible | 03:36 | |
| *** m|y|k has joined #openstack-ansible | 03:37 | |
| *** m|y|k has quit IRC | 03:37 | |
| *** thorst has quit IRC | 03:40 | |
| *** woodard_ has quit IRC | 03:53 | |
| *** thorst has joined #openstack-ansible | 04:12 | |
| *** thorst has quit IRC | 04:17 | |
| *** michelv has joined #openstack-ansible | 04:21 | |
| cloudnull | and https://review.openstack.org/#/c/523642/ -- looks like the handler needed to be fixed up. | 04:21 |
| *** gkadam_ has joined #openstack-ansible | 04:23 | |
| *** threestrands has quit IRC | 04:29 | |
| *** phalmos has joined #openstack-ansible | 04:30 | |
| *** markvoelker has quit IRC | 04:30 | |
| *** armaan has joined #openstack-ansible | 04:37 | |
| *** threestrands has joined #openstack-ansible | 04:41 | |
| *** threestrands has quit IRC | 04:41 | |
| *** threestrands has joined #openstack-ansible | 04:41 | |
| *** gouthamr has quit IRC | 04:43 | |
| *** thorst has joined #openstack-ansible | 04:44 | |
| *** thorst has quit IRC | 04:49 | |
| *** nshetty has joined #openstack-ansible | 04:52 | |
| *** vnogin has joined #openstack-ansible | 04:52 | |
| *** vnogin has quit IRC | 04:57 | |
| *** woodard has joined #openstack-ansible | 04:57 | |
| *** poopcat has quit IRC | 05:07 | |
| *** thorst has joined #openstack-ansible | 05:21 | |
| *** nshetty is now known as nshetty|afk | 05:23 | |
| *** thorst has quit IRC | 05:26 | |
| *** thorst has joined #openstack-ansible | 05:29 | |
| *** markvoelker has joined #openstack-ansible | 05:30 | |
| *** threestrands has quit IRC | 05:32 | |
| *** thorst has quit IRC | 05:34 | |
| *** hybridpollo has quit IRC | 05:35 | |
| *** phalmos has quit IRC | 05:35 | |
| *** chyka has joined #openstack-ansible | 05:36 | |
| *** chyka has quit IRC | 05:41 | |
| *** nshetty|afk is now known as nshetty | 05:43 | |
| *** gouthamr has joined #openstack-ansible | 05:47 | |
| *** thorst has joined #openstack-ansible | 05:57 | |
| *** thorst has quit IRC | 06:03 | |
| *** gouthamr has quit IRC | 06:04 | |
| *** udesale has joined #openstack-ansible | 06:10 | |
| *** thorst has joined #openstack-ansible | 06:29 | |
| *** thorst has quit IRC | 06:34 | |
| *** michelv has quit IRC | 06:37 | |
| *** nwonknu has joined #openstack-ansible | 06:48 | |
| *** esberglu has quit IRC | 06:54 | |
| *** gkadam_ has quit IRC | 06:58 | |
| *** ThomasS has quit IRC | 06:58 | |
| *** armaan has quit IRC | 07:03 | |
| *** thorst has joined #openstack-ansible | 07:07 | |
| *** pmannidi has quit IRC | 07:07 | |
| *** thorst has quit IRC | 07:11 | |
| *** michelv has joined #openstack-ansible | 07:15 | |
| *** Oku_OS-away is now known as Oku_OS | 07:15 | |
| *** armaan has joined #openstack-ansible | 07:19 | |
| *** michelv has quit IRC | 07:20 | |
| *** nshetty is now known as nshetty|brb | 07:22 | |
| *** mbuil has joined #openstack-ansible | 07:36 | |
| *** nshetty|brb has quit IRC | 07:36 | |
| *** nshetty has joined #openstack-ansible | 07:37 | |
| *** thorst has joined #openstack-ansible | 07:42 | |
| *** thorst has quit IRC | 07:46 | |
| *** taseer1 has joined #openstack-ansible | 07:54 | |
| *** Taseer has quit IRC | 07:55 | |
| *** nshetty is now known as nshetty|lunch | 07:55 | |
| *** taseer2 has joined #openstack-ansible | 07:55 | |
| *** shardy has joined #openstack-ansible | 07:56 | |
| *** taseer2 is now known as Taseer | 07:59 | |
| *** taseer1 has quit IRC | 08:00 | |
| *** pcaruana has joined #openstack-ansible | 08:04 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts stable/pike: tasks: lxc_net: Do not source non-existing network script on CentOS https://review.openstack.org/522660 | 08:10 |
| *** ashak has joined #openstack-ansible | 08:13 | |
| *** tnogisto2 has quit IRC | 08:13 | |
| *** thorst has joined #openstack-ansible | 08:21 | |
| *** gkadam_ has joined #openstack-ansible | 08:22 | |
| *** thorst has quit IRC | 08:26 | |
| *** vnogin has joined #openstack-ansible | 08:29 | |
| *** vnogin has quit IRC | 08:34 | |
| jafeha | good morning | 08:34 |
| evrardjp | good morning jafeha | 08:37 |
| evrardjp | odyssey4me: hwoarang andymccr do you remember how many ppl were in our room at the PTG? Around 12 I guess? | 08:38 |
| odyssey4me | evrardjp I guess 10-15 or so | 08:38 |
| evrardjp | Next question, how much time to we need together? I think that 3 days were good. It's from 0.5 to 3D | 08:39 |
| evrardjp | I'd have loved asking those questions yesterday at the community meeting, if I'd have had the questions on time :) | 08:39 |
| evrardjp | Let's say we have the same attendance? | 08:40 |
| *** epalper has joined #openstack-ansible | 08:40 | |
| evrardjp | odyssey4me: I don't know if you've seen but https://review.openstack.org/#/c/404735/ merged | 08:44 |
| *** tnogisto has joined #openstack-ansible | 08:45 | |
| odyssey4me | evrardjp yep, saw that - exciting stuff... no someone has to implement the horizon changes to take advantage of it ! | 08:47 |
| odyssey4me | evrardjp I think that 3D are good. We've done 1D and it wasn't enough. 2D was OK. 3D has been the best to do both the design discussions and getting down to do some work. | 08:48 |
| odyssey4me | typically day 1 is broad, day 2 gets into some details and day 3 is productive | 08:48 |
| evrardjp | yeah that's EXACTLY what I thought | 08:51 |
| evrardjp | On the attendance, I don't think we'll get more ppl because it's in europe, it will probably balance the US ppl not coming. | 08:52 |
| *** thorst has joined #openstack-ansible | 08:54 | |
| *** Taseer has quit IRC | 08:56 | |
| *** thorst has quit IRC | 08:59 | |
| *** nshetty|lunch is now known as nshetty | 09:02 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-galera_server master: Update Percona XtraDB Backup version https://review.openstack.org/523074 | 09:05 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_congress master: Introduce os_congress role in gerrit https://review.openstack.org/522491 | 09:06 |
| *** mwe1 has quit IRC | 09:09 | |
| *** mwe1 has joined #openstack-ansible | 09:10 | |
| *** Taseer has joined #openstack-ansible | 09:11 | |
| *** armaan has quit IRC | 09:17 | |
| *** loudgefly has joined #openstack-ansible | 09:18 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-pip_install master: Add find-links back to pip.conf https://review.openstack.org/523793 | 09:18 |
| hwoarang | ptg yeah! toot toot | 09:19 |
| evrardjp | hwoarang: :) | 09:21 |
| evrardjp | odyssey4me: great! | 09:21 |
| evrardjp | hwoarang: could you vote on odyssey4me 's patch righ there ^^^ | 09:22 |
| evrardjp | please? | 09:22 |
| loudgefly | Hi, I recently deployed a pike infrastructure using delliscsi driver as cinder backend (3 infra nodes, as metal configuration). Volume creation seems to work, but I'm concerned about understanding how cinder instantiates iscsi connections... | 09:22 |
| loudgefly | Iscisiadm shows no sessions | 09:22 |
| loudgefly | nor multipath gives me any hint | 09:23 |
| *** thorst has joined #openstack-ansible | 09:24 | |
| evrardjp | loudgefly: I am not familiar with this driver, maybe you should ask the question in #openstack-cinder ? | 09:25 |
| evrardjp | loudgefly: would you mind sharing the documentation for it in our cinder role? | 09:25 |
| odyssey4me | loudgefly the connections will be direct from the hypervisor to the storage | 09:25 |
| evrardjp | odyssey4me: but shouldn't it be listed there? | 09:25 |
| odyssey4me | cinder-volume only facilitates the creation of the volume and the transfer of the appropriate information in order to do setup/teardown | 09:26 |
| odyssey4me | so, it will matter where it is that you actually check for the connections | 09:26 |
| evrardjp | I mean if it's using iscsi, I'd expect to see the at least mpath devices appearing somewhere in the logs | 09:26 |
| odyssey4me | ie the control plane will not show you much | 09:26 |
| evrardjp | oh yeah I assume it was checked on the compute node/libvirt | 09:27 |
| *** yifei has quit IRC | 09:27 | |
| loudgefly | Ah ok, now It's a little bit clearer: I created a volume but never used it, so I can't check anything on the compute nodes.... | 09:28 |
| evrardjp | loudgefly: :) | 09:28 |
| loudgefly | I should use/mount it and check paths on compute nodes | 09:28 |
| loudgefly | Thanks a lot guys, great work btw | 09:29 |
| evrardjp | loudgefly: yeah right now you've created something onto your dell infra, maybe you can check the volume appears in the SAN management interface | 09:29 |
| *** thorst has quit IRC | 09:29 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update variable scopes https://review.openstack.org/523468 | 09:31 |
| *** jonher_ has joined #openstack-ansible | 09:35 | |
| *** hw_wutianwei has quit IRC | 09:36 | |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_congress master: Introduce os_congress role in gerrit https://review.openstack.org/522491 | 09:37 |
| *** jonher has quit IRC | 09:37 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server master: SUSE: Apply workaround for mariadb-10.2 https://review.openstack.org/523430 | 09:41 |
| *** pbandark has joined #openstack-ansible | 09:43 | |
| *** armaan has joined #openstack-ansible | 09:46 | |
| *** vnogin has joined #openstack-ansible | 09:48 | |
| openstackgerrit | Adrien Cunin proposed openstack/openstack-ansible-ops master: Added ability to cleanup venv tgz https://review.openstack.org/523827 | 09:50 |
| evrardjp | gerrit is slow today :( | 09:50 |
| openstackgerrit | Merged openstack/openstack-ansible-pip_install master: Add find-links back to pip.conf https://review.openstack.org/523793 | 09:51 |
| *** vnogin has quit IRC | 09:53 | |
| *** thorst has joined #openstack-ansible | 09:55 | |
| *** loudgefly_ has joined #openstack-ansible | 09:55 | |
| *** vnogin has joined #openstack-ansible | 09:55 | |
| *** loudgefly has quit IRC | 09:58 | |
| *** thorst has quit IRC | 09:59 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create stable/pike: Fix container restart handler https://review.openstack.org/523832 | 10:02 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-galera_server stable/pike: SUSE: Apply workaround for mariadb-10.2 https://review.openstack.org/523833 | 10:02 |
| openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_congress master: Introduce os_congress role in gerrit https://review.openstack.org/522491 | 10:02 |
| evrardjp | Adri2000: are you there? | 10:04 |
| evrardjp | Adri2000: this broke the translation build site: https://github.com/openstack/openstack-ansible-os_horizon/blob/master/tasks/horizon_post_install.yml#L87-L88 for a while, I didn't notice. | 10:05 |
| evrardjp | I don't have the local_settings.d on my build, are you sure it deserves to be there? | 10:06 |
| *** nshetty is now known as nshetty|afk | 10:10 | |
| Adri2000 | evrardjp: I know it fixes the issue I had. also I just checked I do have other files there: | 10:14 |
| Adri2000 | # ls /openstack/venvs/horizon-16.0.3/lib/python2.7/site-packages/openstack_dashboard/local/local_settings.d/ | 10:14 |
| Adri2000 | _10_set_custom_theme.py.example _12_toggle_data_upload_max_number_fields.py _20_integration_tests_scaffolds.py.example | 10:14 |
| Adri2000 | _11_toggle_angular_features.py.example _2010_integration_tests_deprecated.py.example _9030_profiler_settings.py.example | 10:14 |
| evrardjp | I don't have that folder | 10:20 |
| evrardjp | Adri2000: api content enabled exceptions.py __init__.py locale test utils.py version.py | 10:20 |
| evrardjp | that's all I have in /openstack/venvs/horizon-master/lib/python2.7/dist-packages/sahara_dashboard | 10:20 |
| evrardjp | oh wait. We are not talking about the same path, that's the issue. | 10:20 |
| evrardjp | where do you get your sahara file? | 10:22 |
| *** electrofelix has joined #openstack-ansible | 10:22 | |
| evrardjp | Adri2000: I meant that file https://github.com/openstack/openstack-ansible-os_horizon/blob/master/tasks/horizon_post_install.yml#L87 | 10:22 |
| evrardjp | I am not sure this task is needed, because I have the local/local_settings.d/_12_ blabla without linking it from sahara. | 10:23 |
| *** nshetty|afk is now known as nshetty | 10:25 | |
| Adri2000 | evrardjp: it should already be in the sahara_dashboard path, as it's provided by sahara-dashboard. but it's going to be in use only if we actually copy/link it into the openstack_dashboard path | 10:25 |
| Adri2000 | are you saying that without this task, you already have the file in the openstack_dashboard path? :o | 10:26 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Flexible horizon dashboard configuration https://review.openstack.org/523838 | 10:28 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Flexible horizon dashboard configuration https://review.openstack.org/523838 | 10:28 |
| evrardjp | Adri2000: yes | 10:29 |
| evrardjp | and I don't have this file in sahara-dashboard | 10:29 |
| evrardjp | let me double check if that's normal or not | 10:29 |
| evrardjp | that is very weird. | 10:30 |
| *** thorst has joined #openstack-ansible | 10:31 | |
| mardim | evrardjp, hello | 10:31 |
| mardim | evrardjp, I am trying to load my kerenl module | 10:31 |
| mardim | evrardjp, But it is failing here https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/configure_metal_hosts.yml#L61 | 10:32 |
| mardim | evrardjp, with error dict object' has no attribute 'pattern' | 10:32 |
| mardim | evrardjp, I do not know what to set to the variable pattern | 10:32 |
| mardim | evrardjp, Because I am installing my own kernel modules | 10:33 |
| mardim | evrardjp, I know the name but not the pattern | 10:33 |
| mardim | evrardjp, is this a bug ? | 10:34 |
| mardim | evrardjp, Because at the task below(61 line) | 10:34 |
| mardim | evrardjp, You are setting the pattern key optional | 10:34 |
| mardim | kernel* | 10:34 |
| evrardjp | Adri2000: found the issue. | 10:35 |
| evrardjp | will dig deeper into the root cause. | 10:35 |
| evrardjp | mardim: you can't load a kernel module without a pattern | 10:35 |
| *** thorst has quit IRC | 10:36 | |
| evrardjp | you have to tell what's the module to load, from something you find in config. | 10:36 |
| mardim | evrardjp, My jernel module is not in config | 10:36 |
| mardim | evrardjp, I am installing it with a package and then I need to load it | 10:36 |
| evrardjp | mardim: how can it load? | 10:36 |
| mardim | evrardjp, Just with the name | 10:36 |
| evrardjp | maybe we should have a bypass then | 10:37 |
| mardim | evrardjp, I already done that in the past | 10:37 |
| mardim | and it is workuing | 10:37 |
| evrardjp | but that's supposed to be a safety check | 10:37 |
| mardim | what safety check ? | 10:37 |
| evrardjp | not try to load a module if that's not a module present in config | 10:37 |
| evrardjp | you can write a patch to allow a manual override | 10:38 |
| mardim | evrardjp, Some modules are not necessarly in the config | 10:38 |
| mardim | evrardjp, Maybe the user want to load his own | 10:38 |
| evrardjp | or alternatively you can do a lineinfile for the config | 10:38 |
| mardim | if I add patern: {{}} | 10:38 |
| evrardjp | mardim: they probably know what they are doing and probably need an override. | 10:38 |
| mardim | it will work or ? | 10:38 |
| evrardjp | I don't know I have to check the code... | 10:39 |
| mardim | let me check that | 10:39 |
| evrardjp | try :p | 10:39 |
| evrardjp | :) | 10:39 |
| mardim | ok :) | 10:39 |
| *** loudgefly_ has quit IRC | 10:40 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server master: Implement a proper WSREP check for galera https://review.openstack.org/520665 | 10:41 |
| mardim | evrardjp, Doesn't work I just checked that :( | 10:43 |
| mardim | And I was so close :( | 10:43 |
| *** stuartgr has joined #openstack-ansible | 10:44 | |
| mardim | evrardjp, Do you think that we should remove that task entirely ? | 10:44 |
| evrardjp | mardim: that task work fine | 10:44 |
| evrardjp | it always worked in the past | 10:45 |
| evrardjp | you should just have an override | 10:45 |
| evrardjp | I will check that later, could you ping me in like 30 minutes? | 10:45 |
| mardim | evrardjp, what you mean an override can you explaini more :) | 10:45 |
| mardim | sure I will | 10:45 |
| mardim | thanks :) | 10:45 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Flexible horizon dashboard configuration https://review.openstack.org/523838 | 10:46 |
| evrardjp | gerrit is dead, gerrit is dead! | 10:47 |
| evrardjp | that sounded more like a good tune in my head | 10:48 |
| *** loudgefly_ has joined #openstack-ansible | 10:49 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_horizon master: Flexible horizon dashboard configuration https://review.openstack.org/523838 | 10:49 |
| *** holmsten has joined #openstack-ansible | 10:54 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update variable scopes https://review.openstack.org/523468 | 10:55 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update variable scopes https://review.openstack.org/523468 | 10:57 |
| *** Taseer has quit IRC | 11:04 | |
| *** thorst has joined #openstack-ansible | 11:08 | |
| *** thorst has quit IRC | 11:13 | |
| *** Taseer has joined #openstack-ansible | 11:20 | |
| *** mrch has joined #openstack-ansible | 11:22 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/pike: tasks: lxc_net: Do not source non-existing network script on CentOS https://review.openstack.org/522660 | 11:29 |
| *** thorst has joined #openstack-ansible | 11:30 | |
| *** thorst has quit IRC | 11:32 | |
| *** armaan has quit IRC | 11:42 | |
| *** armaan has joined #openstack-ansible | 11:42 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_server master: Simplify check address whitelisting https://review.openstack.org/523850 | 11:54 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update all SHAs https://review.openstack.org/522850 | 11:55 |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_tacker master: Provide support for openSUSE https://review.openstack.org/523852 | 11:57 |
| *** chyka has joined #openstack-ansible | 11:58 | |
| evrardjp | odyssey4me: it's broken right now. | 12:02 |
| evrardjp | so we need to do something for galera server. | 12:02 |
| odyssey4me | evrardjp what's broken right now? | 12:02 |
| evrardjp | either we rollback to use 0.0.0.0/0 for xinetd | 12:02 |
| evrardjp | or we adopt my kind of patch | 12:02 |
| *** chyka has quit IRC | 12:02 | |
| evrardjp | I can remove haproxy_all if you prefer | 12:03 |
| *** thorst has joined #openstack-ansible | 12:03 | |
| odyssey4me | not sure why we're trying to make it complicated in the role - why not just set it to something basic in the role, or open entirely - then override in group_vars using the inventory? | 12:03 |
| evrardjp | ok let me rollback to 0.0.0.0/0 | 12:04 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_server master: Simplify check address whitelisting https://review.openstack.org/523850 | 12:05 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_server master: Simplify check address whitelisting https://review.openstack.org/523850 | 12:06 |
| *** Tahvok has joined #openstack-ansible | 12:06 | |
| odyssey4me | evrardjp suggestion in PS2 | 12:07 |
| *** thorst has quit IRC | 12:07 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Harden the xinetd security https://review.openstack.org/523854 | 12:08 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_server master: Simplify check address whitelisting https://review.openstack.org/523850 | 12:09 |
| evrardjp | odyssey4me: done | 12:09 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update all SHAs https://review.openstack.org/522850 | 12:11 |
| odyssey4me | LGTM | 12:11 |
| *** smatzek has joined #openstack-ansible | 12:11 | |
| evrardjp | did the group vars implementation too | 12:12 |
| evrardjp | same topic | 12:12 |
| *** pylot has quit IRC | 12:16 | |
| *** armaan has quit IRC | 12:16 | |
| *** armaan_ has joined #openstack-ansible | 12:16 | |
| mardim | evrardjp, Is it now a good time to see what we can do with the kernel module issue or ? | 12:16 |
| *** pylot has joined #openstack-ansible | 12:18 | |
| *** thorst has joined #openstack-ansible | 12:18 | |
| *** thorst has quit IRC | 12:24 | |
| evrardjp | mardim: so your problem is not at fail fast if we can't load a module | 12:30 |
| evrardjp | right? | 12:30 |
| evrardjp | mardim: does it fail when undefined? Because the condition should be there: https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/configure_metal_hosts.yml#L67 | 12:32 |
| *** Tahvok has quit IRC | 12:32 | |
| *** armaan_ has quit IRC | 12:32 | |
| mardim | evrardjp, Sorry wrong link | 12:32 |
| mardim | evrardjp, Let me paste the correct link | 12:32 |
| mardim | evrardjp, it fails here https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/configure_metal_hosts.yml#L54 | 12:33 |
| openstackgerrit | Merged openstack/openstack-ansible-plugins master: Preserve comments in INI files https://review.openstack.org/522557 | 12:33 |
| mardim | evrardjp, the problem is at fail fat because it cannot find the pattern | 12:33 |
| evrardjp | mardim: I don't think so | 12:34 |
| mardim | fat -> fast* | 12:34 |
| mardim | evrardjp, I am seeing the failutre in ansible logs | 12:34 |
| evrardjp | mardim: I think it fails because it finds the pattern | 12:34 |
| evrardjp | != -1 | 12:34 |
| openstack | evrardjp: Error: "=" is not a valid command. | 12:34 |
| evrardjp | != -1 | 12:34 |
| evrardjp | not == -1 | 12:35 |
| mardim | evrardjp, let me paste all the error log so we can be in the same page | 12:35 |
| mardim | evrardjp, https://hastebin.com/pezolagesa.md | 12:36 |
| mardim | here ^ | 12:36 |
| evrardjp | so in this case you defined an empty pattern | 12:36 |
| evrardjp | so what's the issue when not having an empty pattern, or having a pattern that's relevant to your use case? | 12:37 |
| mardim | evrardjp, exactly because I do not have any pattern | 12:37 |
| mardim | evrardjp, So ithought that an empty pattern will solve the issue | 12:37 |
| evrardjp | what if you don't define the pattern then? | 12:37 |
| mardim | again error | 12:37 |
| mardim | let me paste it | 12:37 |
| evrardjp | the empty pattern makes it worse | 12:37 |
| evrardjp | :p | 12:37 |
| mardim | evrardjp, https://hastebin.com/yahugiboxo.md | 12:38 |
| mardim | evrardjp, Yes I think the same :) | 12:38 |
| *** armaan has joined #openstack-ansible | 12:38 | |
| evrardjp | mardim: could you add a condition https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/configure_metal_hosts.yml#L59 ? | 12:39 |
| evrardjp | on your host | 12:39 |
| evrardjp | - item.pattern is defined | 12:39 |
| mardim | evrardjp, Not sure what you mean here can you please explain a bit more :) | 12:40 |
| *** jwitko has quit IRC | 12:40 | |
| evrardjp | mardim: on your node where you are testing | 12:40 |
| evrardjp | edit your /etc/ansible/roles/openstack_hosts/tasks/configure_metal_hosts.yml | 12:41 |
| evrardjp | add a line at L59 | 12:41 |
| evrardjp | - item.pattern is defined | 12:41 |
| evrardjp | and then test again | 12:41 |
| mardim | a ok yes I will do | 12:41 |
| *** threestrands has joined #openstack-ansible | 12:42 | |
| *** threestrands has quit IRC | 12:42 | |
| *** threestrands has joined #openstack-ansible | 12:42 | |
| *** zkynet has joined #openstack-ansible | 12:47 | |
| mardim | evrardjp, Do you want to test that by incuding the pattern key at the user_variables.yaml or ? | 12:47 |
| zkynet | hey guys, I'm trying to deploy openstack-ansible but the combo of python + ansible that is installed via the boostrap-ansible.sh script is broken | 12:48 |
| evrardjp | mardim: no pattern | 12:48 |
| evrardjp | this way it's undefined | 12:48 |
| evrardjp | but we should probably add tests into the role indeed | 12:48 |
| evrardjp | both cases if not done already | 12:48 |
| *** armaan has quit IRC | 12:48 | |
| mardim | ok I am running lets see | 12:48 |
| zkynet | unsupported operand type(s) for %: 'bytes' and 'bytes' ... anyone seen this before ? | 12:48 |
| evrardjp | zkynet: that sounds weird | 12:49 |
| evrardjp | we are using python2 by default | 12:49 |
| *** armaan has joined #openstack-ansible | 12:49 | |
| evrardjp | zkynet: it should work with python2 if you follow the procedure and I haven't seen a gate issue, so I am surprised | 12:49 |
| evrardjp | what's your environment? | 12:49 |
| zkynet | I'm deploying from the ocata docs | 12:49 |
| evrardjp | ok ocata, not master | 12:50 |
| evrardjp | let me check | 12:50 |
| zkynet | vagrant box: ubuntu/trusty64 | 12:50 |
| zkynet | I think it's at 14.04 | 12:50 |
| mardim | evrardjp, https://hastebin.com/sifijixate.sql | 12:50 |
| mardim | failed ^^ | 12:50 |
| evrardjp | zkynet: haha | 12:51 |
| evrardjp | yeah | 12:51 |
| evrardjp | that's not supported by anyone | 12:51 |
| evrardjp | :) | 12:51 |
| evrardjp | the maximum openstack version supported on 14.04 is newton | 12:52 |
| evrardjp | you should upgrade to 16.04 | 12:52 |
| zkynet | so I need 16.04 ? | 12:52 |
| zkynet | roger ! | 12:52 |
| zkynet | thanks ^^ | 12:52 |
| evrardjp | zkynet: https://docs.openstack.org/project-deploy-guide/openstack-ansible/ocata/overview-requirements.html#software-requirements | 12:52 |
| zkynet | I'll report back if the problem resolves it self from the update | 12:52 |
| evrardjp | zkynet: our latest build of today worked on Ocata. | 12:53 |
| evrardjp | 16.04 | 12:53 |
| zkynet | how did I miss that :( .. for some reason I thought it was 14.04 that was supported | 12:53 |
| zkynet | thank you <3 | 12:53 |
| evrardjp | no worries :) | 12:53 |
| evrardjp | I am glad we aren't broken :p | 12:53 |
| evrardjp | we generally aren't on old stable branches, so I was concerned! | 12:54 |
| *** sboyron has joined #openstack-ansible | 12:54 | |
| *** zkynet has quit IRC | 12:54 | |
| *** thorst has joined #openstack-ansible | 12:55 | |
| *** zkynet_ has joined #openstack-ansible | 12:55 | |
| *** zkynet_ is now known as zkynet | 12:55 | |
| mardim | evrardjp, if is set this like this: | 12:58 |
| mardim | evrardjp, https://hastebin.com/gahakatete.swift | 12:59 |
| mardim | it is working | 12:59 |
| mardim | if is -> if I | 13:01 |
| *** threestrands has quit IRC | 13:01 | |
| evrardjp | great | 13:02 |
| evrardjp | mardim: could you submit a patch there? | 13:02 |
| mardim | with the change that I just did ? | 13:02 |
| evrardjp | Allow the pattern to be undefined | 13:02 |
| mardim | evrardjp, ^ | 13:02 |
| evrardjp | yeah | 13:02 |
| Bico_Fino | Morning. | 13:02 |
| mardim | evrardjp, Sure | 13:02 |
| evrardjp | I can do it too if you prefer | 13:02 |
| evrardjp | Bico_Fino: morning sir! | 13:02 |
| mardim | evrardjp, No I will no worries | 13:03 |
| mardim | :) | 13:03 |
| evrardjp | thanks | 13:03 |
| mardim | no rpoblem thanks for the quick response :) | 13:03 |
| evrardjp | I will just mark myself on it, because I kinda wrote the code that goes with it :) | 13:03 |
| mardim | problem* | 13:03 |
| mardim | evrardjp, You can do whatever you want :) | 13:03 |
| evrardjp | Bico_Fino: how can we help you? | 13:03 |
| Bico_Fino | How do I upgrade the repo containers? Need some packages there for Ubuntu 16.0.3, but only has 16.0.1 | 13:03 |
| Bico_Fino | an openstack-ansible repo-install.yml is enough? :) | 13:04 |
| evrardjp | Bico_Fino: I generally delete and re-create out of laziness | 13:04 |
| evrardjp | but yes, repo_install by default has package state: latest IIRC | 13:04 |
| evrardjp | Bico_Fino: what is the issue you're encountering? | 13:04 |
| Bico_Fino | galera cluster went south | 13:04 |
| Bico_Fino | dunno why yet. | 13:05 |
| Bico_Fino | rsyslog container was full, but idea deploying with Debug: True. hehe | 13:05 |
| evrardjp | maybe worth cleaning things ups, cleaning up logs :) | 13:07 |
| Bico_Fino | If change Debug to False, I need run what? Just setup-infra? | 13:08 |
| evrardjp | debug: is used everywhere :p | 13:09 |
| evrardjp | you're good to run it everywhere :) | 13:09 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Harden the xinetd security https://review.openstack.org/523854 | 13:11 |
| Bico_Fino | 98G Nov 29 13:11 ./log/infra3_neutron_agents_container-baaee0b6/neutron/neutron-l3-agent.log | 13:11 |
| Bico_Fino | 98g | 13:11 |
| Bico_Fino | ouch | 13:12 |
| evrardjp | lol | 13:12 |
| mhayden | buenos dias! | 13:17 |
| evrardjp | goeiemorgend! | 13:18 |
| *** markvoelker has quit IRC | 13:19 | |
| *** markvoelker has joined #openstack-ansible | 13:19 | |
| mhayden | evrardjp: want me to craft a reno for https://review.openstack.org/#/c/523850/ in a follow-up patch? | 13:26 |
| mhayden | i | 13:26 |
| mhayden | i'm thinking about people that are possibly using our galera_server role for their own usage | 13:26 |
| mhayden | they might get surprised by this change | 13:27 |
| mhayden | (as i was) | 13:27 |
| hwoarang | i am still not sure about that patch :/ | 13:27 |
| mhayden | hwoarang: it makes more sense when examined along with 523854 | 13:27 |
| openstackgerrit | Lihi Wishnitzer proposed openstack/openstack-ansible-os_neutron master: Update Draognflow configuration https://review.openstack.org/523870 | 13:27 |
| lihi | evrardjp, ^^ | 13:29 |
| evrardjp | mhayden: there was no release note on talking about the new variables in https://github.com/openstack/openstack-ansible-galera_server/commit/94821f8108f589e47f8c164625f4309ec477a600 | 13:30 |
| evrardjp | lihi: ok | 13:30 |
| hwoarang | mhayden: yeah i understand the reasoning but the 'default' one feels like too risky. people may not notice that the default is too open :) | 13:31 |
| hwoarang | maybe we should fail if the default is unchanged ;p | 13:31 |
| hwoarang | or pick a default that will cause non-safe deployments to fail so people are forced to do the right thing. but anyway :) | 13:32 |
| mgariepy | morning | 13:32 |
| *** udesale has quit IRC | 13:33 | |
| evrardjp | hwoarang: mhayden that was EXACTLY my point at the initial patch, then cloudnull came with the chnage that broke the gates, so I thought, let's simplify. I had a patch, then it was not good for odyssey4me , so I moved to 0.0.0.0/0 | 13:33 |
| evrardjp | and then many table flips followed. | 13:33 |
| evrardjp | :p | 13:33 |
| evrardjp | I just don't care what's the solution, I want it to stop breaking gates. | 13:34 |
| mhayden | trolololol | 13:34 |
| evrardjp | :) | 13:34 |
| mhayden | so worst case scenario is if someone is using our role externally and they don't set the allowed addresses | 13:34 |
| evrardjp | and then open xinetd stuff! | 13:34 |
| odyssey4me | the role should not be tied to the inventory, the inventory stuff should come from group vars | 13:34 |
| mhayden | if we look at this role all by itself, would a setting of 127.0.0.1 break the deployment ? | 13:34 |
| odyssey4me | a reno is definitely needed | 13:35 |
| evrardjp | yeah if we do not have group vars | 13:35 |
| odyssey4me | perhaps the role should not enable this functionality by default, then we enable it through group vars | 13:35 |
| evrardjp | odyssey4me: I misreviewed the first patch. We should have a reno on that patch, if not, a reno should be added as a separate patch | 13:35 |
| odyssey4me | and through group vars, give it the right stuff | 13:35 |
| evrardjp | odyssey4me: sounds a better approach maybe | 13:36 |
| evrardjp | until set, do not do it | 13:36 |
| evrardjp | let me fix that quick | 13:36 |
| odyssey4me | yep | 13:36 |
| odyssey4me | then the open default doesn't matter | 13:36 |
| evrardjp | that could work with a reno too, because that would be a new feature | 13:36 |
| evrardjp | :) | 13:36 |
| hwoarang | odyssey4me: what does "the role should not be tied to the inventory, the inventory stuff should come from group vars" mean? | 13:36 |
| *** dave-mccowan has joined #openstack-ansible | 13:36 | |
| hwoarang | in the end the role depends on the inventory right? | 13:36 |
| evrardjp | hwoarang: roles should be independant of the inventory it applies to | 13:37 |
| evrardjp | you should apply a role to inventory nodes, not the other way around | 13:37 |
| evrardjp | Imagine if we add all roles running on 'all' hosts and we add conditionals, what a mess it would be! | 13:37 |
| evrardjp | :) | 13:37 |
| odyssey4me | this pattern hasn't been applied all through our roles, but it's been a goal to move towards it for a long time | 13:37 |
| hwoarang | i see | 13:38 |
| evrardjp | odyssey4me: hwoarang mhayden I will refactor the https://review.openstack.org/#/c/523850/ then | 13:38 |
| hwoarang | right | 13:39 |
| *** hw_wutianwei has joined #openstack-ansible | 13:42 | |
| *** openstackstatus has quit IRC | 13:43 | |
| *** openstack has joined #openstack-ansible | 13:46 | |
| *** ChanServ sets mode: +o openstack | 13:46 | |
| mardim | mbuil, https://review.openstack.org/#/c/523875/ | 13:46 |
| evrardjp | mhayden: to help on the decision, if I don't set this, by default xinetd is open to everyone :) | 13:46 |
| mhayden | oh, i know :| | 13:47 |
| mhayden | it's always in the hardening guides | 13:47 |
| mhayden | although i think it was pulled from the RHEL 7 STIG | 13:47 |
| mhayden | based on the assumption that nobody uses xinetd any more | 13:47 |
| evrardjp | so no-access is not defined? | 13:47 |
| evrardjp | ok will do | 13:47 |
| evrardjp | :p | 13:47 |
| *** sboyron has quit IRC | 13:50 | |
| *** loudgefly_ has quit IRC | 13:51 | |
| *** loudgefly has joined #openstack-ansible | 13:51 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create stable/pike: Fix container restart handler https://review.openstack.org/523832 | 13:52 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-galera_server master: Fix galera_monitoring_allowed_source https://review.openstack.org/523850 | 13:52 |
| *** sboyron has joined #openstack-ansible | 13:53 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Allow the loading of modules with undefined pattern https://review.openstack.org/523875 | 13:54 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Fix the loading of modules with undefined pattern https://review.openstack.org/523875 | 13:58 |
| *** udesale has joined #openstack-ansible | 14:01 | |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-ops master: Validate the target directory exist before linking. https://review.openstack.org/522869 | 14:05 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update variable scopes https://review.openstack.org/523468 | 14:06 |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible-ops master: Validate the target directory exists before linking. https://review.openstack.org/522869 | 14:06 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update variable scopes https://review.openstack.org/523468 | 14:06 |
| mgariepy | mhayden, ^^ | 14:06 |
| mhayden | ORLY | 14:07 |
| mgariepy | this one: https://review.openstack.org/522869 | 14:09 |
| *** dasTor_ has joined #openstack-ansible | 14:10 | |
| *** dasTor has quit IRC | 14:13 | |
| *** shardy has quit IRC | 14:14 | |
| *** woodard has quit IRC | 14:21 | |
| cloudnull | mornings | 14:21 |
| *** nshetty has quit IRC | 14:21 | |
| *** woodard has joined #openstack-ansible | 14:22 | |
| cloudnull | oh, did i break the gate again ... | 14:22 |
| mhayden | so i was looking at nova's pike release notes -> The configuration options baremetal_enabled_filters and use_baremetal_filters are deprecated in Pike and should only be used if your deployment still contains nodes that have not had their resource_class attribute set. | 14:22 |
| mhayden | we're using the use_baremetal_filters right now in os_nova | 14:23 |
| mhayden | i';m not familiar enough with ironic to know if we should remove that | 14:23 |
| mhayden | prometheanfire: ^^ ? | 14:23 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server master: Tune-up the galera role for efficiency https://review.openstack.org/466827 | 14:26 |
| cloudnull | evrardjp: odyssey4me: do I need to add a reno somewhere? | 14:27 |
| mgariepy | cloudnull, can you review https://review.openstack.org/#/c/522869/ please ? | 14:28 |
| cloudnull | mgariepy: done. | 14:29 |
| mgariepy | thanks | 14:29 |
| *** smatzek has quit IRC | 14:29 | |
| *** smatzek has joined #openstack-ansible | 14:30 | |
| evrardjp | cloudnull: no it's merged | 14:31 |
| *** esberglu has joined #openstack-ansible | 14:31 | |
| *** smatzek has quit IRC | 14:34 | |
| *** sboyron has quit IRC | 14:38 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_tacker master: Provide support for openSUSE https://review.openstack.org/523852 | 14:42 |
| *** jwitko has joined #openstack-ansible | 14:46 | |
| *** rguichard has joined #openstack-ansible | 14:46 | |
| *** sboyron has joined #openstack-ansible | 14:46 | |
| *** openstack has quit IRC | 14:46 | |
| *** openstack has joined #openstack-ansible | 14:52 | |
| *** ChanServ sets mode: +o openstack | 14:52 | |
| *** hw_wutianwei has quit IRC | 14:52 | |
| cloudnull | hwoarang: question, if you're around. | 14:53 |
| *** dasTor_ has quit IRC | 14:54 | |
| hwoarang | cloudnull: yep | 14:54 |
| *** dasTor has joined #openstack-ansible | 14:54 | |
| *** sxc731 has joined #openstack-ansible | 14:54 | |
| cloudnull | so this is what i was working on https://review.openstack.org/#/c/523218/ which needed newer systemd and only because the systemd-netword service / tools are not available in leap | 14:54 |
| cloudnull | that also has a tie to the nspawn bits I'm working on as well | 14:54 |
| cloudnull | as I'd like to use networkd for inner container networking | 14:55 |
| cloudnull | so the question is, do you know of a package to get those bits I can install? | 14:55 |
| cloudnull | or is that something I can build in a non-impactful way ? | 14:56 |
| hwoarang | let me check | 14:57 |
| *** yifei has joined #openstack-ansible | 14:57 | |
| *** rguichard has quit IRC | 14:58 | |
| hwoarang | so cloudnull you are basically missing networkd | 14:58 |
| cloudnull | yup | 14:58 |
| hwoarang | systemd-networkd | 14:58 |
| hwoarang | that's explicitly disabled on the leap build. let me find out why | 14:58 |
| hwoarang | i need to source the right person :) | 14:58 |
| cloudnull | ++ | 14:58 |
| cloudnull | thanks ! | 14:58 |
| cloudnull | like in centos there's an additional package I can install to enable it | 14:59 |
| cloudnull | if there's something like that we can use that'd be fine/great | 14:59 |
| sxc731 | Hi team, I have a super-easy q for those in the know (no doubt): I'm trying to define some neutron ML2 config (eg: neutron_provider_networks) so they differ between agents and server containers. | 14:59 |
| sxc731 | I thought I might be able to put them in /etc/openst...deploy/group_vars/neutron_{agent,server}.yml but this doesn't seem to get picked up? | 14:59 |
| cloudnull | I can continue to carry on and simply case "when: zypper" etc, but having it all be the same would be sweet! IMHO. | 15:00 |
| cloudnull | sxc731: so you want different config for different nodes? | 15:00 |
| *** Neptu has joined #openstack-ansible | 15:00 | |
| sxc731 | For different container groups really | 15:01 |
| *** armaan has joined #openstack-ansible | 15:01 | |
| cloudnull | is the split agent/server? | 15:01 |
| sxc731 | I see no reason for neutron_server containers to have access to the data plane (eg: VLAN networks) for example, unless I'm missingsmth? | 15:01 |
| cloudnull | ah. | 15:02 |
| cloudnull | so you can change that in the openstack_user_config file . | 15:02 |
| sxc731 | Agents and server land in different containers which have different NIC bindings) | 15:02 |
| cloudnull | under the provider_networks there's a "group_bind" | 15:02 |
| cloudnull | set that accordingly | 15:02 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.aio#L51 | 15:03 |
| sxc731 | Yes, I got the network config in the containers the way I want; the problem is that ml2_conf.ini cannot refer to things like "tenant_network_types = vlan" (in the server containers, unlike agent), else it doesn't come up | 15:04 |
| *** savvas_ has joined #openstack-ansible | 15:05 | |
| cloudnull | hum. | 15:06 |
| *** marst has joined #openstack-ansible | 15:06 | |
| sxc731 | (hopefully this makes some sense). So the variables I'm trying to customize - based on container group - are things like neutron_provider_networks (these are normally set in user_variables.yml) | 15:06 |
| cloudnull | so the neutron_provider_networks variable comes in here: https://github.com/openstack/openstack-ansible/blob/master/playbooks/common-playbooks/neutron.yml#L83-L92 | 15:07 |
| hwoarang | cloudnull: i found more info but i will have to email people to see if it's possible to get it on Leap. if not then we will see. | 15:07 |
| cloudnull | hwoarang: thanks hwoarang! | 15:08 |
| cloudnull | if it becomes to big a deal, no worries, I can work around it :) | 15:08 |
| cloudnull | and if its something we'll have to wait for leap 15, so be it. | 15:08 |
| cloudnull | I know your busy, but if its something we can get/enable easily that'd be lovely. | 15:10 |
| hwoarang | i can't see it in leap-15 either | 15:10 |
| hwoarang | so let me find out more info | 15:10 |
| cloudnull | sxc731: looks like we need an option to detect the presence of neutron_provider_networks and use that when defined. | 15:10 |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Run dstat in gate tests https://review.openstack.org/522799 | 15:10 |
| cloudnull | ^ woot! | 15:10 |
| sxc731 | cloudnull: many thanks for your help so far! So no way this can be defined on a per container-group-basis currently? I kind of thought this is what "group_vars" are for but admittedly not much of an Ansible expert ;-) | 15:12 |
| *** smatzek has joined #openstack-ansible | 15:12 | |
| evrardjp | hwoarang: could you revote on https://review.openstack.org/#/c/523850/ ? | 15:13 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Allow neutron_provider_networks to be user defined https://review.openstack.org/523899 | 15:14 |
| cloudnull | sxc731: ^ | 15:14 |
| cloudnull | mind testing that change out? | 15:15 |
| cloudnull | that should give you the ability to define it in group_vars | 15:15 |
| *** woodard has quit IRC | 15:15 | |
| *** yifei has quit IRC | 15:17 | |
| sxc731 | cloudnull: cool, I'll let you know how this fares in a few mins! | 15:19 |
| *** sboyron has quit IRC | 15:19 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update all SHAs https://review.openstack.org/522850 | 15:20 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Update variable scopes https://review.openstack.org/523468 | 15:20 |
| *** gouthamr has joined #openstack-ansible | 15:22 | |
| *** sboyron has joined #openstack-ansible | 15:26 | |
| *** yifei has joined #openstack-ansible | 15:27 | |
| *** tlbr has joined #openstack-ansible | 15:28 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/ocata: Run dstat in gate tests https://review.openstack.org/523902 | 15:30 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible stable/pike: scripts: scripts-library.sh: Use pgrep -f to find the dstat process https://review.openstack.org/523904 | 15:31 |
| openstackgerrit | Periyasamy Palanisamy proposed openstack/openstack-ansible master: <wip> install quagga for configuring ODL for BGP https://review.openstack.org/523907 | 15:32 |
| *** yifei has quit IRC | 15:33 | |
| openstackgerrit | Periyasamy Palanisamy proposed openstack/openstack-ansible master: <wip> install quagga for configuring ODL for BGP https://review.openstack.org/523907 | 15:34 |
| *** Tahvok has joined #openstack-ansible | 15:35 | |
| openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_tacker master: Provide support for openSUSE https://review.openstack.org/523852 | 15:37 |
| prometheanfire | mhayden: those filters should be removed | 15:37 |
| prometheanfire | keep in mind they it said depricated, not absent | 15:37 |
| mgariepy | does anyone knows it the nova-placement-api needs to be reached from the internet ? | 15:38 |
| mhayden | true | 15:38 |
| prometheanfire | I THINK that means they are still around | 15:38 |
| mgariepy | i can't seems to be able to find out the info anywhere.. | 15:38 |
| mhayden | mgariepy: i thought the placement api was only consumed internally | 15:38 |
| mhayden | i know that nova-conductor must be able to reach it in pike | 15:38 |
| *** mrch has quit IRC | 15:38 | |
| *** weezS has joined #openstack-ansible | 15:40 | |
| hwoarang | mbuil: do you want to add an opensuse job to https://review.openstack.org/#/c/523852/3 ? | 15:40 |
| mgariepy | ok thanks mhayden | 15:40 |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: Only whitelist internal network for placement api https://review.openstack.org/523913 | 15:42 |
| mbuil | hwoarang: yes, I added it to my backlog of tasks 10 minutes ago :P | 15:43 |
| hwoarang | mbuil: we can wait until you add one in teh patchset to see that it actually works | 15:43 |
| *** tnogisto2 has joined #openstack-ansible | 15:43 | |
| mbuil | hwoarang: I promise you it does ==> https://hastebin.com/muqumonuva.lua :) | 15:44 |
| *** armaan has quit IRC | 15:45 | |
| *** savvas_ has quit IRC | 15:45 | |
| *** armaan has joined #openstack-ansible | 15:45 | |
| mbuil | hwoarang: I have just tried. I am not sure when will I get time to add the opensuse jobs, I am currently a bit overloaded | 15:45 |
| hwoarang | can i edit your patch and add one myself? | 15:45 |
| *** savvas_ has joined #openstack-ansible | 15:45 | |
| *** tnogisto has quit IRC | 15:46 | |
| mbuil | hwoarang: yes, that would be awesome! | 15:46 |
| hwoarang | it's just two lines of code | 15:46 |
| hwoarang | ok lets do it | 15:46 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tacker master: Provide support for openSUSE https://review.openstack.org/523852 | 15:47 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible-os_neutron master: Add OvS-NSH support https://review.openstack.org/517259 | 15:47 |
| sxc731 | cloudnull: seems to work a treat; you're a hero! Don't think I ever had a change request implemented so quickly in my (20 years') career... | 15:51 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_horizon master: Remove enable_{firewall,vpn} options https://review.openstack.org/523916 | 15:52 |
| sxc731 | cloudnull: in order to get this to work (app-level testing pending), I also had to set 'neutron_ml2_drivers_type' in the group_vars files. And I had to soft-link neutron_agent.yml -> compute_all.yml so the neutron-linuxbridge-agent on the compute hosts also gets the config | 15:53 |
| sxc731 | cloudnull: is it worth trying to put some doc together (happy to take a shot at it) or is this too far-fetched a use-case to warrant this? | 15:55 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible master: Fix crontab errors on CentOS/SUSE https://review.openstack.org/520682 | 15:57 |
| cloudnull | sxc731: I think this would be a good thing to doc. | 15:57 |
| cloudnull | having the option and covering what's possible is something I can see folks wanting. | 15:58 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Use Ansible 2.4 https://review.openstack.org/522778 | 16:01 |
| sxc731 | cloudnull:cool, thx again! So stick this somewhere in here? https://docs.openstack.org/openstack-ansible-os_neutron/latest/ I'm making that "generous" offer but haven't done this before (though I at least have appropriate git config to contribute); I'll try to figure out how it's done and come back to ask for some hand-holding if needs be (hopefully not!) | 16:02 |
| mgariepy | any core available to review : https://review.openstack.org/#/c/522869/ | 16:03 |
| cloudnull | sxc731: that looks like a good spot. | 16:03 |
| *** phalmos has joined #openstack-ansible | 16:04 | |
| cloudnull | sxc731: the commit semantics are simple with gerrit, tl;dr make the change, add a good commit message, run `git review` instead of `git push` when you're ready. | 16:04 |
| * cloudnull idk if you've messed with gerrit before, so feel free to tell me to be quite :) | 16:04 | |
| cloudnull | ++ mgariepy 522869 is a good change to get in for leap. | 16:05 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Do not apply varstest to all scenarios https://review.openstack.org/523920 | 16:06 |
| sxc731 | cloudnull: sounds nice & easy ; not had a lot of exposure to Gerrit and certainly wouldn't tell you to be quiet after the amazing help today! | 16:06 |
| cloudnull | thanks for testing and reporting the issue! community for the win! | 16:07 |
| evrardjp | :) | 16:07 |
| *** woodard has joined #openstack-ansible | 16:07 | |
| sxc731 | cloudnull: absolutely! Amazing OpenStack community as ever! | 16:07 |
| evrardjp | sometimes I feel the worst invention on earth is a modern web browser. | 16:07 |
| spotz | sxc731: If youo need instructions for setting up git and gerrit let me know | 16:07 |
| cloudnull | odyssey4me: https://review.openstack.org/#/c/466827 merged! very nice! | 16:08 |
| cloudnull | now to do take aim at rabbitmq | 16:08 |
| sxc731 | spotz: thanks, will do | 16:08 |
| odyssey4me | cloudnull ja, awesome - thanks for taking that on | 16:10 |
| *** sxc731_ has joined #openstack-ansible | 16:10 | |
| cloudnull | anyone around to give this a review https://review.openstack.org/#/c/521229/ now that the tune-up changes are in. | 16:10 |
| * mhayden ganders | 16:10 | |
| * mhayden realizes he has already completed gandering | 16:11 | |
| cloudnull | odyssey4me: i was just building on what you started. | 16:11 |
| cloudnull | ^ it takes a village :D | 16:11 |
| cloudnull | mhayden: ++ thanks for gandering | 16:11 |
| *** sxc731 has quit IRC | 16:13 | |
| *** armaan has quit IRC | 16:13 | |
| *** phalmos has quit IRC | 16:14 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Validate the target directory exists before linking. https://review.openstack.org/522869 | 16:14 |
| *** armaan has joined #openstack-ansible | 16:14 | |
| *** sxc731_ has quit IRC | 16:15 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Remove the AIO scenario & add new scenarios to maintain coverage https://review.openstack.org/516002 | 16:15 |
| *** savvas_ has quit IRC | 16:15 | |
| cloudnull | I'd also appreciate review ^ there too. | 16:15 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: [DNM] Build everything on metal https://review.openstack.org/504224 | 16:19 |
| *** sxc731 has joined #openstack-ansible | 16:19 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Build everything on metal https://review.openstack.org/504224 | 16:19 |
| evrardjp | ^ passes tempest on my machine :) | 16:20 |
| evrardjp | 45 minutes | 16:20 |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts master: Fix the loading of modules with undefined pattern https://review.openstack.org/523875 | 16:20 |
| evrardjp | let's see how it will behave on poor gates | 16:20 |
| *** zkynet has quit IRC | 16:21 | |
| odyssey4me | evrardjp mhayden cloudnull so with https://review.openstack.org/#/c/523850/5 will haproxy not be able to use it - ie it is implemented, but unusable | 16:23 |
| evrardjp | yes | 16:23 |
| evrardjp | that's what you asked! | 16:23 |
| *** zkynet has joined #openstack-ansible | 16:24 | |
| evrardjp | https://review.openstack.org/#/c/523854/2 | 16:24 |
| odyssey4me | alright, so does that mean https://review.openstack.org/520673 needs to be adjusted to enable it? or is there another patch to go on top of it? | 16:24 |
| mhayden | TIL that neutron can have a dns_domain set on a particular port that overrides the main dns_domain | 16:24 |
| evrardjp | odyssey4me: ^ https://review.openstack.org/#/c/523854/2 | 16:24 |
| odyssey4me | well, for it to go on top of actually | 16:24 |
| evrardjp | I have 3 patches depending on 523854 ;) | 16:25 |
| *** sboyron has quit IRC | 16:25 | |
| *** phalmos has joined #openstack-ansible | 16:26 | |
| *** sxc731 has quit IRC | 16:31 | |
| mbuil | hwoarang: the errors I get in the tacker patch are probably related to how the tests are defined or? | 16:37 |
| *** holmsten has quit IRC | 16:44 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Use Ansible 2.4 https://review.openstack.org/522778 | 16:46 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_neutron master: Remove send_arp_for_ha option https://review.openstack.org/523930 | 16:47 |
| *** sxc731 has joined #openstack-ansible | 16:48 | |
| zkynet | @evratdjp: the boostrap process does indeed work with ubuntu 16, even in docker containers ^^ | 16:51 |
| *** zkynet has quit IRC | 16:52 | |
| *** udesale has quit IRC | 17:00 | |
| *** chyka has joined #openstack-ansible | 17:01 | |
| hwoarang | mbuil: i think there is a problem with the galera role | 17:03 |
| hwoarang | let me check | 17:03 |
| *** epalper has quit IRC | 17:11 | |
| *** TxGirlGeek has joined #openstack-ansible | 17:11 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible master: Remove jinja2 delimiters in when https://review.openstack.org/523936 | 17:13 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Use Ansible 2.4 https://review.openstack.org/522778 | 17:16 |
| *** gkadam_ has quit IRC | 17:27 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_neutron master: Change deprecated nova_metadata_ip option https://review.openstack.org/523941 | 17:29 |
| *** savvas_ has joined #openstack-ansible | 17:31 | |
| *** mbuil has quit IRC | 17:33 | |
| *** savvas_ has quit IRC | 17:35 | |
| *** loudgefly has quit IRC | 17:37 | |
| *** vnogin has quit IRC | 17:48 | |
| *** thorst has quit IRC | 18:00 | |
| *** vnogin has joined #openstack-ansible | 18:03 | |
| *** vnogin has quit IRC | 18:04 | |
| *** thorst has joined #openstack-ansible | 18:06 | |
| SamYaple | you guys just merged this without even responding to me :( https://review.openstack.org/#/c/520665/ | 18:08 |
| *** RandomTech has joined #openstack-ansible | 18:09 | |
| odyssey4me | cloudnull ^ | 18:11 |
| *** thorst has quit IRC | 18:11 | |
| cloudnull | SamYaple: I'm actually on it, but yes it was merged this morning. | 18:16 |
| *** michelv has joined #openstack-ansible | 18:16 | |
| SamYaple | cool. just dont want the comment missed. potentially the patch you have in right now has quite a few race conditions with checking that local_state key | 18:17 |
| cloudnull | the current implementation pulls us in line with what percona recommends, I will be following up to make better. | 18:17 |
| SamYaple | oh do they have docs on that? | 18:18 |
| cloudnull | they do | 18:18 |
| *** thorst has joined #openstack-ansible | 18:18 | |
| cloudnull | https://www.percona.com/doc/percona-xtradb-cluster/LATEST/howtos/virt_sandbox.html | 18:18 |
| SamYaple | beautiful! thanks | 18:18 |
| cloudnull | also this outside of the HA reference guide https://www.percona.com/doc/percona-xtradb-cluster/LATEST/howtos/haproxy.html | 18:19 |
| cloudnull | I just implemented their spec | 18:20 |
| *** thorst has quit IRC | 18:22 | |
| SamYaple | got it cloudnull. yea then its far more widely used so that greatly reduces my concerns about potential issues. | 18:26 |
| SamYaple | i think my "what ifs" can be safely ignored | 18:26 |
| cloudnull | I think what you have there should be part of it | 18:26 |
| cloudnull | because if we can ensure the what if's never happen, everyone wins :D | 18:27 |
| SamYaple | the Primary/Non-Primary thing is good for sure | 18:27 |
| SamYaple | really helps with online split-brain type recoveries too | 18:27 |
| cloudnull | ++ | 18:27 |
| SamYaple | given the same wsrep position, you want to bootstrap the node thats running vs the node thats stopped for IST/SST reasons. | 18:28 |
| SamYaple | thats what initially lead me to using it | 18:28 |
| *** armaan has quit IRC | 18:35 | |
| *** gkadam_ has joined #openstack-ansible | 18:35 | |
| RandomTech | hey cloudnull will the playbooks run up if the deployment host is accessing the targets then the container network? | 18:37 |
| cloudnull | mhayden: any good way to get this to use a different repo as a fallback: http://logs.openstack.org/50/523850/5/check/openstack-ansible-functional-centos-7/ed5af3c/job-output.txt.gz#_2017-11-29_17_18_13_664594 | 18:37 |
| cloudnull | looks like copr-be.cloud.fedoraproject.org was telling us to go thank ourselves. | 18:37 |
| cloudnull | RandomTech: what's that? | 18:38 |
| RandomTech | the container network? | 18:38 |
| *** germs has joined #openstack-ansible | 18:38 | |
| cloudnull | mhayden: http://paste.openstack.org/show/627753/ | 18:39 |
| cloudnull | mgariepy: ^ | 18:39 |
| RandomTech | oops i meant a diffrent network than the container network. we had a subnet set asside for specifcally ansible | 18:39 |
| cloudnull | so you want to change the network containers are run on, or add another network to the containers for ansible things ? | 18:40 |
| *** thorst has joined #openstack-ansible | 18:40 | |
| cloudnull | something like seperating ansible deployment traffic from the normal management interface? | 18:40 |
| *** thorst has quit IRC | 18:40 | |
| *** thorst has joined #openstack-ansible | 18:40 | |
| RandomTech | pretty much the second one. will it cause problems? | 18:41 |
| cloudnull | not at all. | 18:41 |
| cloudnull | set this flag https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.aio#L41 | 18:42 |
| cloudnull | to where it needs to be and ansible will run their. | 18:42 |
| cloudnull | **there | 18:42 |
| *** germs1 has joined #openstack-ansible | 18:42 | |
| *** sxc731 has quit IRC | 18:43 | |
| *** cshen_ has joined #openstack-ansible | 18:43 | |
| cloudnull | RandomTech: you might have something like this | 18:43 |
| cloudnull | http://paste.openstack.org/show/627754/ | 18:44 |
| *** germs has quit IRC | 18:44 | |
| cloudnull | the management traffic would run over br-mgmt and the deployment traffic over br-deploy | 18:44 |
| cloudnull | if that's what you're looking for. | 18:45 |
| *** michelv has quit IRC | 18:48 | |
| *** michelv has joined #openstack-ansible | 18:48 | |
| *** cshen_ has quit IRC | 18:49 | |
| *** cshen_ has joined #openstack-ansible | 18:51 | |
| jrosser | cloudnull: as the containers are accessed via the host ive approached that deploy-net thing by treating the "bare metal" interface as the one all the ansible-ing happens over | 18:53 |
| jrosser | no needto bring that into the containers? | 18:53 |
| cloudnull | ++ | 18:54 |
| cloudnull | that works perfectly well. | 18:54 |
| cloudnull | the only exception I can think of is an adhoc ansible command which may not be using our plugins. | 18:54 |
| cloudnull | but come to think of it, i may be over thinking it :) | 18:55 |
| jrosser | ive gone so far as having all the bare metal interfaces as native vlan and them in their own vrf | 18:56 |
| jrosser | they cant grt to br-mgmt even if they try directly | 18:56 |
| cloudnull | that's a cool idea ! | 18:56 |
| *** TxGirlGeek has quit IRC | 18:57 | |
| jrosser | you have to delegate_to somewhere thats actually supposed to be able to | 18:57 |
| *** sxc731 has joined #openstack-ansible | 18:59 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:00 | |
| *** sleeveace has joined #openstack-ansible | 19:00 | |
| *** poopcat has joined #openstack-ansible | 19:00 | |
| mgariepy | cloudnull, that's sad.. | 19:03 |
| mgariepy | copr stuff. | 19:03 |
| *** poopcat1 has joined #openstack-ansible | 19:03 | |
| *** electrofelix has quit IRC | 19:04 | |
| *** poopcat has quit IRC | 19:05 | |
| *** Oku_OS is now known as Oku_OS-away | 19:06 | |
| cloudnull | I want to say i've seen this before. | 19:08 |
| cloudnull | so i suspect it's rebuilding or something | 19:08 |
| cloudnull | but would be nice to have an alternative | 19:08 |
| *** armaan has joined #openstack-ansible | 19:15 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible master: Remove jinja2 delimiters in when https://review.openstack.org/523936 | 19:16 |
| *** armaan has quit IRC | 19:17 | |
| mhayden | jmccrory: ^^ should be fixed up | 19:17 |
| *** armaan has joined #openstack-ansible | 19:17 | |
| *** poopcat1 is now known as poopcat | 19:18 | |
| evrardjp | thanks mhayden | 19:19 |
| *** stuartgr has quit IRC | 19:19 | |
| evrardjp | I still do these mistakes myself, even if I know them too well. | 19:20 |
| *** cshen_ has quit IRC | 19:21 | |
| *** cshen_ has joined #openstack-ansible | 19:22 | |
| mgariepy | cloudnull, you could host a mirror ? | 19:22 |
| *** michelv has quit IRC | 19:29 | |
| *** TxGirlGeek has quit IRC | 19:29 | |
| evrardjp | cloudnull: when is x rechecks too much? for me that was 15. | 19:30 |
| *** TxGirlGeek has joined #openstack-ansible | 19:31 | |
| *** cshen_ has quit IRC | 19:32 | |
| *** smatzek has quit IRC | 19:32 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Use Ansible 2.4 https://review.openstack.org/522778 | 19:33 |
| *** cshen_ has joined #openstack-ansible | 19:34 | |
| *** smatzek has joined #openstack-ansible | 19:38 | |
| *** sleeveace has quit IRC | 19:39 | |
| *** smatzek_ has joined #openstack-ansible | 19:41 | |
| *** smatzek has quit IRC | 19:43 | |
| *** cshen_ has quit IRC | 19:45 | |
| cloudnull | looks like it's back | 19:45 |
| cloudnull | http://copr-be.cloud.fedoraproject.org/ | 19:45 |
| *** smatzek_ has quit IRC | 19:46 | |
| *** smatzek has joined #openstack-ansible | 19:46 | |
| *** hamza has joined #openstack-ansible | 19:47 | |
| *** TxGirlGeek has quit IRC | 19:51 | |
| mhayden | oh did COPR go offline? | 19:51 |
| cloudnull | yup | 19:51 |
| *** smatzek_ has joined #openstack-ansible | 19:51 | |
| *** smatzek has quit IRC | 19:51 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:52 | |
| mhayden | :| | 19:52 |
| *** cshen_ has joined #openstack-ansible | 19:53 | |
| cloudnull | folks thoughts on https://review.openstack.org/#/c/523525/ | 19:57 |
| cloudnull | evrardjp: ^ | 19:57 |
| *** TxGirlGeek has quit IRC | 20:00 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Gather more logs when running tests https://review.openstack.org/523647 | 20:02 |
| *** hybridpollo has joined #openstack-ansible | 20:06 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Harden the xinetd security https://review.openstack.org/523854 | 20:09 |
| *** hybridpollo has quit IRC | 20:14 | |
| RandomTech | hey does the playbook update a dns table somewhere? | 20:15 |
| cloudnull | the openstack-hosts role updates the /etc/hosts file. | 20:16 |
| cloudnull | it will append the file. | 20:16 |
| cloudnull | we also have an unbound playbook | 20:17 |
| cloudnull | which IMHO is a better general approach to internal dns | 20:17 |
| RandomTech | do you know where something like 'ctrl-01.openstack.local' would be defined? | 20:17 |
| RandomTech | were trying to ping an adress and its try to look for that instead | 20:18 |
| cloudnull | the "openstack.local" part is defined within the role | 20:18 |
| cloudnull | the ctrl-01 part is set in the openstack_user_config | 20:18 |
| cloudnull | for you host. | 20:18 |
| RandomTech | im now seeing that in the host file strangly | 20:18 |
| RandomTech | not* | 20:19 |
| SamYaple | the place im at now has '.' in thier hostname, so the fqdn is 'host.name'.domain.tld | 20:19 |
| SamYaple | which breaks all dns things everywhere | 20:19 |
| cloudnull | SamYaple: :\ | 20:19 |
| RandomTech | for some strange reason my deployment host cant reach the target hosts on the container network but they can all reach eachother | 20:20 |
| cloudnull | different subnets or vlans? | 20:21 |
| RandomTech | and when i try to ping the adress manually it pings for 'ctrl-01.openstack.local' instead | 20:21 |
| *** smatzek_ has quit IRC | 20:21 | |
| RandomTech | they apear to be on the same subnet and vlan to me | 20:21 |
| cloudnull | you could remove the items in /etc/hosts | 20:22 |
| RandomTech | the target host can see the request using tcpdump | 20:22 |
| cloudnull | and see if the problem persists ? | 20:22 |
| *** smatzek has joined #openstack-ansible | 20:22 | |
| cloudnull | maybe the routing table needs to be updaated? | 20:22 |
| RandomTech | i dont see the adress in the hosts file however | 20:23 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-tests master: Gather more logs when running tests https://review.openstack.org/523647 | 20:23 |
| *** smatzek has quit IRC | 20:24 | |
| RandomTech | its really strange | 20:25 |
| cloudnull | maybe there's more than one host with ctrl-01 on your network ? | 20:26 |
| *** smatzek has joined #openstack-ansible | 20:26 | |
| cloudnull | ** with the name ctrl-01 ** | 20:26 |
| RandomTech | its the one im trying to reach using an ip '10.45.0.2' | 20:28 |
| *** smatzek has quit IRC | 20:28 | |
| cloudnull | so you're ping'ing 10.45.0.2 and its sending you elsewhere? | 20:28 |
| RandomTech | and when i tcpdump i get an arp request for 'ctrl-01.openstack.local' from '10.45.0.1' (the pinging machine) | 20:28 |
| cloudnull | like arp who-has? | 20:28 |
| RandomTech | ya | 20:29 |
| RandomTech | ctrl-01 is seeing this | 20:29 |
| RandomTech | which is strange | 20:29 |
| *** smatzek has joined #openstack-ansible | 20:29 | |
| *** smatzek has quit IRC | 20:30 | |
| RandomTech | apparently the target host has that adress defined, but i have no clue how the deloying host is getting it | 20:31 |
| cloudnull | is it an alias on your deploy host? | 20:32 |
| *** smatzek_ has joined #openstack-ansible | 20:33 | |
| *** smatzek__ has joined #openstack-ansible | 20:35 | |
| *** smatzek__ has quit IRC | 20:35 | |
| *** germs1 is now known as germs | 20:35 | |
| *** smatzek__ has joined #openstack-ansible | 20:35 | |
| *** hybridpollo has joined #openstack-ansible | 20:36 | |
| *** smatzek_ has quit IRC | 20:37 | |
| RandomTech | im not seeing anything it /etc/hosts but it is definately changing it somewhere as when i ping 10.45.0.55 it looks for the proper ip | 20:38 |
| *** sxc731_ has joined #openstack-ansible | 20:41 | |
| RandomTech | is there anywhere on the deployment host it updates soem hosts names? | 20:43 |
| cloudnull | it should never change the hostname | 20:43 |
| cloudnull | it will add to the hosts file | 20:43 |
| cloudnull | but not to the /etc/hostname | 20:43 |
| *** sxc731_ has quit IRC | 20:44 | |
| *** foutatoro has joined #openstack-ansible | 20:44 | |
| RandomTech | ya my /etc/hosts file is looking like this: http://paste.openstack.org/show/627761/ | 20:46 |
| RandomTech | the 45 adress is nowhere int here | 20:46 |
| RandomTech | or is there a seperate hosts file im missing | 20:47 |
| *** threestrands has joined #openstack-ansible | 20:47 | |
| bndzor | kernel:[335796.132002] BUG: soft lockup - CPU#5 stuck for 22s! [swapper/5:0]/code | 20:50 |
| bndzor | anyone got an idea whats causuing this ? | 20:50 |
| cloudnull | on centos ? | 20:50 |
| bndzor | seems to be occuring on one of the vms running on one of the compute nodes | 20:50 |
| bndzor | debian 8 | 20:50 |
| cloudnull | is there a traceback ? | 20:51 |
| cloudnull | RandomTech: there's just the one host file | 20:51 |
| bndzor | was all i had time to see before everything froze | 20:51 |
| *** hamza has quit IRC | 20:52 | |
| bndzor | moved the vm to another compute node to see if it occurs agian | 20:52 |
| cloudnull | RandomTech: our host file additions will only cover the management network 10.45.0.1 | 20:52 |
| cloudnull | ** is 10.45.0.1 part of management? | 20:52 |
| cloudnull | bndzor: let me know if you figure anything out. | 20:53 |
| cloudnull | could just be a kernel bug in debian 8 ? | 20:53 |
| bndzor | on the official image from deb ? | 20:53 |
| bndzor | would be pretty well known by now then i guess :P | 20:54 |
| RandomTech | it was supposed to be however we ran the playbook with specifying all the host using a management only ip. not the propper container network ip | 20:54 |
| cloudnull | I would hope so. | 20:54 |
| RandomTech | and ansible only ip* | 20:54 |
| cloudnull | can you just rerun the openstack-host-setup playbook ? | 20:55 |
| cloudnull | now that you have the ips fixed ? | 20:55 |
| cloudnull | or is there something blocking that? | 20:55 |
| RandomTech | it cant reach them | 20:55 |
| RandomTech | which is what im trying to fix | 20:55 |
| cloudnull | this sounds like a routing issue | 20:56 |
| RandomTech | This is how the interfaces look: http://paste.openstack.org/show/627762/ | 20:57 |
| RandomTech | and tcpdump proves there seeing trafic from eachother | 20:58 |
| cloudnull | what does `ip route` show you ? | 20:58 |
| RandomTech | ctrl-01 has this 10.45.0.0/16 dev br-mgmt proto kernel scope link src 10.45.0.2 | 20:59 |
| RandomTech | ansible has this: 10.45.0.0/16 dev enp4s0 proto kernel scope link src 10.45.0.1 | 20:59 |
| cloudnull | any other routes that might be conflicting? | 21:00 |
| *** pcaruana has quit IRC | 21:00 | |
| *** TxGirlGeek has joined #openstack-ansible | 21:02 | |
| *** cshen_ has quit IRC | 21:05 | |
| *** cshen_ has joined #openstack-ansible | 21:06 | |
| SamYaple | so cloudnull, as ive dicovered in the last ~3 months or so and pinned down the behaviour, galera has quorum. good. but when a node cleanly shuts down it *leaves* quorum. so you can go from 5 nodes down to 1 node, and that one node will happily accept writes | 21:08 |
| SamYaple | that kinda breaks the "ha" agreement a bit, writes in one location | 21:09 |
| SamYaple | so you might want to enforce a "only allow writes if >50% of backends are active" | 21:09 |
| SamYaple | maybe | 21:09 |
| SamYaple | anyway just a thought, you might already be aware | 21:09 |
| sxc731 | jamesdenton: hi and sorry to barge in, could I ask a quick q on neutron-server ML2 config? | 21:12 |
| openstackgerrit | Merged openstack/openstack-ansible-galera_server stable/pike: SUSE: Apply workaround for mariadb-10.2 https://review.openstack.org/523833 | 21:15 |
| jamesdenton | hi sxc731 - sure | 21:17 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: Upgrade to RabbitMQ 3.6.14 https://review.openstack.org/522865 | 21:17 |
| sxc731 | Given OSA deploys neutron-server in containers separate from agents, I'm trying to figure out the "correct"/minimal way to configure ml2_conf.ini. I assume it needs to know about the configured tenant network types (such as VLAN), as well as some related meta-data (such as segmentation IDs). However, it should _not_ need to know any details of the physical mapping of such networks (linux-bridge-agent cares, but it's not coll | 21:18 |
| sxc731 | ocated) | 21:18 |
| *** smatzek__ has quit IRC | 21:18 | |
| jamesdenton | OSA should handle all of that for you. | 21:19 |
| jamesdenton | IIRC, neutron-server doesn't care about the mapping since it doesn't apply, but it will need to know the provider label and vlan mappings | 21:19 |
| sxc731 | hmm... depends. I'm trying to avoid exposing my data plane network interfaces on the neutron-server containers | 21:19 |
| jamesdenton | neutron-server container itself should only have the default nat network and container ip | 21:21 |
| jamesdenton | no data plane interface plumbed | 21:21 |
| jamesdenton | it may be in the config file, since the API needs to validate user input and make things happen, but no actual physical link is needed there | 21:22 |
| sxc731 | Agreed, that's exactly what I'm trying to achieve. The issue is with "network_vlan_ranges = physnet1:1100:1199" under [ml2_type_vlan] in ml2_conf.ini | 21:23 |
| jamesdenton | k | 21:23 |
| sxc731 | Unless there's a mapping for physnet1, neutron-server dies | 21:23 |
| sxc731 | and, per above, we're tryting to avoid exposing physnet1 (the data plane network) | 21:23 |
| jamesdenton | So you need network_vlan_ranges = physnet1, at a minimum | 21:24 |
| jamesdenton | But only the agents themselves need the mapping of physnet1 to eth1, or whatever | 21:24 |
| sxc731 | Ah! you mean w/o the segmentation IDs (in neutron-server's ml2_conf.ini)? | 21:25 |
| jamesdenton | And you're not really exposing the data plane interface itself, rather, you're telling neutron that when a user creates a provider network in the api, that what they're feeding in is legitimate. Or, if its creating a network automatically, which provider network it should use. | 21:25 |
| jamesdenton | so the segmentation IDs are consumed when you don't specify one on the net-create command, or if your tenant network type is vlan and a regular user creates a network. That range is IDs is assigned until they're gone, then the user gets an error | 21:26 |
| sxc731 | Understood. I'm trying to get neutron-server to come up w/o seeing the underlying data plane network for tenants' private VLANs | 21:27 |
| sxc731 | With just 'network_vlan_ranges = physnet1', neutron-server fails with "UnmappedInstanceError: Class 'neutron.objects.plugins.ml2.vlanallocation.VlanAllocation' is not mapped" | 21:28 |
| jamesdenton | Right. neutron-server, the service, won't see any data plane traffic. And the data plane traffic is not exposed at all to the container by OSA. You're simply configuring all known provider network names and respective vlans available on that provider. | 21:28 |
| jamesdenton | hmm. one sec | 21:28 |
| sxc731 | That's exactly what I'm trying to achieve! | 21:29 |
| jamesdenton | what version is this? | 21:31 |
| sxc731 | master. FYI, earlier in the day, cloudnull tweaked OSA so it will now allow configuring neutron-server's ml2_conf.ini with specific config, which is way cool. | 21:32 |
| sxc731 | Now I'm trying to figure out what that config should be and I get the feeling you can't configure neutron-server to know about VLAN tenant_network_types unless you also give it access to the related data plane (unless I'm failing to configure neutron-server's ml2_conf in the right way...) | 21:32 |
| jamesdenton | naw | 21:32 |
| sxc731 | I'm sure I'm missing smth obvious ;-) | 21:33 |
| jamesdenton | You're basically telling Neutron what network types are available to tenants. Meaning, when a tenant does 'neutron net-create testnet', that network *type* is determined by that list, in order of preference | 21:33 |
| sxc731 | Sure. My current issue is I'm not able to bring up neutron-server (this isn't an OSA issue; it's more that the container segmentation afforded by OSA exposes this "difficulty") | 21:35 |
| jamesdenton | I'm on Newton at the moment, and `network_vlan_ranges = physnet1` seems to work OK in this environment. With that config, I'm basically saying ADMIN can create provider networks using the physnet1 'label', and specify any VLAN ID they want. TENANT cannot create a network (of type vlan) because there are no VLAN ID ranges listed on that provider. Make sense? | 21:37 |
| sxc731 | http://paste.openstack.org/show/627763/ is neutron-server's ml2_conf.ini. Fails w/ UnmappedInstanceError: Class 'neutron.objects.plugins.ml2.vlanallocation.VlanAllocation' is not mapped | 21:37 |
| jamesdenton | Yeah, it definitely is a little more difficult since the individual services only require a subset of the configuration options in a given file | 21:37 |
| jamesdenton | k, that looks like a good config. | 21:38 |
| jamesdenton | that's failing for you? | 21:38 |
| sxc731 | Let me paste the neutron-server log... | 21:38 |
| jamesdenton | So, flat is also listed as a type but the block is not there | 21:39 |
| jamesdenton | Try removing flat from tenant_network_types and restart | 21:39 |
| jamesdenton | oh wait nevermind | 21:39 |
| jamesdenton | it's a type_driver. nevermind | 21:39 |
| sxc731 | Sure I'll give that a shot; I though I could have the flat driver but not provide it for tenants to create nets? | 21:39 |
| sxc731 | Ah ok | 21:39 |
| jamesdenton | send that log over | 21:40 |
| sxc731 | http://paste.openstack.org/show/627764/ | 21:40 |
| jamesdenton | Is http://paste.openstack.org/show/627763/ the latest config file? | 21:41 |
| sxc731 | is it perhaps 'mechanism_drivers = linuxbridge'? n-server shouldn't care about that right? | 21:41 |
| sxc731 | naah, same thing | 21:42 |
| sxc731 | This is Pike btw | 21:42 |
| bndzor | cloudnull: it occured again, on another compute node https://pastebin.com/9w9vmgAM | 21:45 |
| bndzor | il give it a newer kernel etc, but still | 21:46 |
| sxc731 | jamesdenton: http://paste.openstack.org/show/627763/ is the latest config file; I have tried with and without ":1100:1199" as suffix to network_vlan_ranges; fails in the same way | 21:47 |
| *** cshen_ has quit IRC | 21:48 | |
| jamesdenton | interesting. I'm on a little older master (pike) and that's working ok. Try re-running the 'openstack-ansible os-neutron-install.yml --tags neutron-config' command to have it revert the config files | 21:49 |
| sxc731 | I'm running pike too. I know it'll fail if I reconfigure because I have used cloudnull's tweak to allow differentiated config for neutron-server... I'm trying to work out the "correct" way to configure neutron-server in the scenario where it doesn't have a data plane interface | 21:52 |
| jamesdenton | What is the tweak? | 21:52 |
| jamesdenton | I'm still trying to figure out how neutron-server gets access to a dataplane interface, sorry | 21:52 |
| sxc731 | cloudnull's tweak is here: https://review.openstack.org/#/c/523899/ It allows using group_vars to configure neutron-server's ml2_conf.ini differently from agent container's and compute nodes' | 21:53 |
| *** weezS has quit IRC | 21:54 | |
| jamesdenton | In what way are you wanting to make them different | 21:56 |
| *** foutatoro has quit IRC | 21:56 | |
| jamesdenton | Like, what would you make your override look like | 21:57 |
| sxc731 | Agents have network_mappings: "neutron_provider_networks.physnet1:eth11", which isn't appropriate for neutron-server as it doesn't have that eth11 data plane interface | 21:58 |
| RandomTech | we figured it out cloudnull the problem was that it was on a trunked port without the traffic being tagged by ubuntu | 22:00 |
| RandomTech | cya | 22:00 |
| *** RandomTech has quit IRC | 22:05 | |
| jamesdenton | sxc731 I see. That provider_network in openstack_user_config.json only is applied to the neutron_openvswitch_agent group. Thus, only the neutron-agent containers have any sort of mapping in their ml2_conf.ini | 22:06 |
| jamesdenton | the neutron server container ml2 file should be absent of a providerlabel:interface mapping. period. | 22:06 |
| jamesdenton | You still need network_vlan_ranges, if you've enabled the vlan type driver | 22:07 |
| sxc731 | Agreed. And cloudnull's tweak allows this (and any other flexibility we might require here) | 22:07 |
| sxc731 | Are you saying you're able to bring up a pike neutron-server with a ml2_conf.ini like this: http://paste.openstack.org/show/627766/ ? | 22:07 |
| jamesdenton | Yes. My ml2_conf.ini looks like that. In Newton. In Pike. In Kilo. | 22:08 |
| jamesdenton | You need the VLAN mappings if you want neutron to automatically allocate vlan IDs on net-create | 22:09 |
| jamesdenton | at a minimum, you need 'network_vlan_ranges = physnet1' | 22:09 |
| sxc731 | Sure, so that's the previous config: http://paste.openstack.org/show/627763/ . In both cases, neutron-server crashes with Unrecoverable error: please check log for details.: UnmappedInstanceError: Class 'n | 22:09 |
| sxc731 | eutron.objects.plugins.ml2.vlanallocation.VlanAllocation' is not mapped | 22:09 |
| jamesdenton | The fact that you're getting an error with that, or with vlans, is concerning but isn't a configuration issue as far as i can tell | 22:10 |
| jamesdenton | yeah I would need to spin up a new install and test. it could be something else is broken | 22:10 |
| jamesdenton | sxc731 Here are the ml2 and ovs config files from my install: http://paste.openstack.org/show/627768/ | 22:12 |
| jamesdenton | the ml2_conf.ini wouldn't dictate, at all, the interface mapping. that's done in the linuxbridge or ovs files | 22:13 |
| jamesdenton | the agent files, rather | 22:13 |
| sxc731 | Ah, you're running ovs, not linux-bridge... could that be the difference? | 22:13 |
| jamesdenton | naw | 22:13 |
| jamesdenton | can you paste: /openstack/venvs/neutron-*/lib/python2.7/site-packages/neutron/conf/plugins/ml2/drivers/driver_type.py | 22:15 |
| sxc731 | http://paste.openstack.org/show/627769/ | 22:16 |
| bndzor | cloudnull: upgraded to 4.9 kernel, nagging about the serial etc is gone.. maybe was just old virtio drivers etc that caused mayhem | 22:18 |
| *** thorst has quit IRC | 22:19 | |
| *** woodard has quit IRC | 22:20 | |
| *** jwitko_ has joined #openstack-ansible | 22:20 | |
| jamesdenton | Not sure what's going on there, sxc731. The config is allowed though, and called out in Line 74 of that file. Something else is causing this issue and I'm not sure what it is | 22:23 |
| *** jwitko has quit IRC | 22:23 | |
| *** jwitko_ has quit IRC | 22:24 | |
| sxc731 | Yes, I see - so it shouldn't make a difference whether I use "<physical_network>:<vlan_min>:<vlan_max>" or just "<physical_network>". And indeed both fail in the same way... | 22:26 |
| sxc731 | Let me paste a more complete stack trace (my previous effort wasn't great ;-) | 22:26 |
| sxc731 | Here goes: http://paste.openstack.org/show/627772/ | 22:34 |
| *** pmannidi has joined #openstack-ansible | 22:36 | |
| sxc731 | jamesdenton: I think I've got it... I must have some crap in my ml2_vlan_allocations table (or smth to that extent). I'm really sorry I might have wasted your time barking up the wrong tree... | 22:43 |
| *** jwitko has joined #openstack-ansible | 22:47 | |
| *** thorst has joined #openstack-ansible | 22:50 | |
| sxc731 | jamesdenton: wow, that was it! I had some data in there with an different (previous test) physnet id and the net result is newton-server crashes if it can't find it.... Deleting the data form the DB did it! | 22:51 |
| jamesdenton | right on! | 22:53 |
| jamesdenton | Glad you fixed it. Not sure really need that patch cloudnull baked up for you, but good luck with it | 22:53 |
| *** thorst has quit IRC | 22:55 | |
| *** esberglu has quit IRC | 22:55 | |
| *** thorst has joined #openstack-ansible | 22:56 | |
| sxc731 | I understand what you're saying now (it was my confusion about this dataplane on neutron-server container stuff). And if I just need different interface name mappings for that dataplane interface between neutron-agent containers and agents running on the compute nodes ("metal"), then that's what "host_bind_override" in openstack_user_config's provider_networks is for, right? | 22:57 |
| *** thorst has quit IRC | 23:01 | |
| *** vnogin has joined #openstack-ansible | 23:03 | |
| *** rromans has quit IRC | 23:11 | |
| *** jonher has joined #openstack-ansible | 23:14 | |
| *** jonher_ has quit IRC | 23:17 | |
| *** TxGirlGeek has quit IRC | 23:19 | |
| *** TxGirlGeek has joined #openstack-ansible | 23:20 | |
| *** TxGirlGeek has quit IRC | 23:24 | |
| *** pbandark has quit IRC | 23:26 | |
| *** sxc731 has quit IRC | 23:26 | |
| *** thorst has joined #openstack-ansible | 23:35 | |
| *** TxGirlGeek has joined #openstack-ansible | 23:38 | |
| *** TxGirlGeek has quit IRC | 23:40 | |
| *** thorst has quit IRC | 23:41 | |
| *** TxGirlGeek has joined #openstack-ansible | 23:48 | |
| *** marst has quit IRC | 23:49 | |
| *** TxGirlGeek has quit IRC | 23:52 | |
| *** TxGirlGe_ has joined #openstack-ansible | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!